Add goldenEvalCekCatchBudget and change existing cases to use it

[SeungheonOh]

goldenEvalCekCatchBudget combines goldenBudget and goldenEvalCekCatch. Unlike calling two separately, this will only run CEK once.

Previous cases that used goldenBudget will use goldenEvalCekCatchBudget since it's more sensible to test their evaluation result as well.

Also, changed some cases to use goldenBundle'. Looks like these were missed out from #7057.

Closes #6846

[effectfully]

Less noise is always a good thing.

[ana-pantilie]

@SeungheonOh I added the "no changelog required" label. It's fine to add it to make CI happy if your changes aren't user facing (like in this case).

[SeungheonOh]

I don't have permission to add labels nor merge anything(even with CI passing and approved review!) yet. For now, please feel free to merge this whenever CI's happy.

[zliu41]

Good work 👍

[zliu41]

I think it's better to create one file containing both budget and result, instead of two?

[effectfully]

I think so too. Budget should go first in case the result is large (can happen for early Plutus versions of if error message is long).

[SeungheonOh]

My reasoning here is based on this comment by @Unisay : #6846 (comment)

Having two files will give better feedback: one can easily check if budgeting is broken or evaluator is broken or both are broken from file changes without going over each files

[zliu41]

I disagree - you don't need to go over the entire file if the budget changed and the result didn't - the diff would clearly show that, especially if the budget appears before the result like @effectfully said.

Having both in a single file makes it easier to review the change (which benefits both the author and the reviewers): when the budget changes, I always want to verify whether the result changed too. Putting them in separate files means I'd need to look for the corresponding .eval file, which may not be adjacent to the .budget file, or included in the PR. With a single file, I can scan through changes much more efficiently.

[zliu41]

There are some golden files that need to be removed, e.g. checkScriptContext2-20.budget.golden. You can delete all golden files and then re-run the tests.

[zliu41]

Try

cabal test plutus-tx-plugin --test-options=--accept
cabal test plutus-benchmark --test-options=--accept
cabal test plutus-ledger-api:plutus-ledger-api-plugin-test --test-options=--accept
cabal test cardano-constitution --test-options=--accept

This should fix all the golden files.